Year: 2026
Type: preprint
Abstract: Earth's energy imbalance has increased rapidly since continuous satellite observations began in 2001, raising concern that equilibrium climate sensitivity may be higher than in recent assessments. Here we use energy-budget trends over 2006–2024 to estimate climate sensitivity in a Bayesian framework, accounting for uncertainty in observations, forcing, variability, and pattern effects. Climate sensitivity inference depends strongly on the likelihood method, and we propose a simple, revised method that is physically consistent with the global energy-balance model. Its maximum likelihood is at 5.2K, whereas the existing method maximizes at 3.0K. Using the revised likelihood and a standard uniform-feedback prior, the posterior median is 3.6K; posterior medians range from 3.3–4.8K depending on the prior. Posterior 66% ranges largely overlap existing assessments of 2–5K and strongly disfavor climate sensitivity below 2K, but upper bounds are prior-dependent. Thus 21st-century trends do not require unusually high climate sensitivity compared to other lines of evidence.
DOI: 10.31223/x5fn68
Author: Vincent Cooper
Institution: Massachusetts Institute of Technology
Language: English
Subfield: Global and Planetary Change
Field: Environmental Science
Domain: Physical Sciences
SDG: Climate action
Open Access status: green
Funder: National Science Foundation
OpenAlex ID: W7214559744